

23·
17 hours agoThe funny thing is that one of the reports on the glasswing project found out a lot (~55%) of the vulnerabilities AI finds are just false positives and not really worth reporting so they just increase the cognitive workload of unpaid maintainers.
The numbers are based on what happened to curl, since they are transparent about their results, and the ones that weren’t false positives were underwhelming low severity vulnerabilities.
Here’s the report: https://www.vulncheck.com/blog/anthropic-glasswing-november-update